Dynamic Computation Offloading for Platooning Vehicles: A Stochastic Game-Theoretic Approach

Edge computing enables mobile users to process their applications by leveraging the computing resource of an edge server nearby and thus receives increasing attention from both academia and industry. It has been considered a promising technique for practically deploying of autonomous systems that require processing massive sensor data, such as autonomous vehicle platoons. In this paper, we investigate the problem of computation offloading in an edge computing-enabled and platoon-based system where a group of platooning vehicles need to make decisions on computation offloading. In particular, we formulate a game-theoretical model regarding the dynamics of the platooning vehicles contending to access wireless channels. The game model aims to maximize each vehicle’s payoff, characterized by energy consumption and total latency in task processing. We propose a distributed stochastic learning algorithm to solve the game model and prove its guaranteed convergence to a pure Nash Equilibrium strategy. We also provide simulation results to validate our algorithm and show its effectiveness in fair resource allocation.